In recent years, Cu(In,Ga)Se2 (CIGS) materials have been extensively studied for use as an absorber layer in thin film photovoltaic devices, owing to their band gaps that can be tuned by adjusting the elemental ratios and are well matched with the solar spectrum (1.1 eV for CuInSe2 to 1.7 eV for CuGaSe2), offering potentially high conversion efficiencies; 20.3% conversion efficiency was achieved using Cu(InxGa1-x)Se2 material by researchers at ZSW and the Centre for Solar Energy and Hydrogen Research in Germany (August 2010). One drawback of CIGS materials is the high manufacturing cost, due to the high cost of the constituent elements. Cu2ZnSnS4 (CZTS) materials can be used as a low-cost alternative to traditional Cu(In,Ga)Se2, due to the abundance and low toxicity of Zn and Sn, which are much cheaper than Ga and the rarer In.
Recently there have been efforts to investigate the direct band gap of this material. CZTS is reported to have a band gap between 1.45 and 1.6 eV [H. Katagiri et al., Appl. Phys. Express, 2008, 1, 041201; K. Ito et al., Jpn. J. Appl. Phys., 1988, 27 (Part 1), 2094; T. M. Friedlmeier et al., Proc. 14th European PVSEC, Barcelona, Spain, 30 Jun. 1997, p. 1242] and a high optical absorption coefficient (up to 105 cm−1) [G. S. Babu et al., J. Phys. D: Appl. Phys., 2008, 41, 205305], which are similar to those of CuInGaSe2. The current record conversion efficiency for pure Cu2ZnSnS4 of 8.4% [B. Shin et al., Prog. Photovolt.: Res. Appl., 2013, 21, 72] shows great potential for this material.
Related compounds, where Zn is partially or completely substituted with another d-block element, Sn is substituted with a group 14 element and/or S is partially or entirely replaced with another chalcogen are also of interest. Examples include Cu2ZnSnSe4, Cu2ZnSnTe4, Cu2CdSnSe4, Cu2CdSnTe4, [H. Matsushita et al., J. Mater. Sci., 2005, 40, 2003] Cu2FeSnS4, [X. Zhang et al., Chem. Commun., 2012, 48, 4656], and Cu2+xZn1-xGeSe4. [W. G. Zeier et al., J. Am. Chem. Soc., 2012, 134, 7147] With the exception of Cu2CoSiSe4, and Cu2NiSiSe4, all compounds in the series Cu2-II-IV(S,Se)4 (II=Mn, Fe, Co, Ni, Zn, Cd, Hg; IV=Si, Ge, Sn) have been synthesised and structurally characterised in the bulk form, as described by Schafer and Nitsche. [W. Schafer and R. Nitsche, Mat. Res. Bull., 1974, 9, 645] Though the substitution of Zn, Sn and/or S may not necessarily offer financial advantages, these compounds are nevertheless attractive for optoelectronic applications as they offer a range of thermodynamic, optical and electrical properties that can be exploited. For example, Cu2Zn1-xCdxSn(Se1-ySy)4 is a semiconductor with a band gap that can be tuned from 0.77 eV (x=0.5, y=0) to 1.45 eV (x=0, y=1) and displays p-type conductivity. [M. Altosaar et al., Phys. Stat. Sol. (a), 2008, 205, 167] Ab initio studies on Cu2ZnSnSe4, Cu2CdSnSe4 and Cu2ZnSnS4 suggest that their thermoelectric performance can be enhanced by Cu doping at the Zn/Cd site. [C. Sevik and T. Ça{hacek over (g)}in in, Phys. Rev. B, 2010, 82, 045202] Other studies have focussed on manipulation of the crystallographic phase of these materials [X. Zhang et al., Chem. Commun., 2012, 48, 4656] and the resultant changes in their electronic properties. [W. Zalewski et al., J. Alloys Compd., 2010, 492, 35]
Methods to produce CIGS- and CZTS-type solar cells with high power conversion efficiency (PCE) often employ vacuum-based deposition of the absorber layer. Vacuum-based approaches typically offer high uniformity, which translates to a high quality film. However, the techniques are also generally costly, with material consumption and energy usage being high. Non-vacuum-based approaches are attractive in that they are typically higher throughput processes, with a lower deposition cost. One such method is a nanoparticle-based deposition approach. Nanoparticles offer several advantages over bulk materials for thin film optoelectronic applications. Firstly, a small amount of nanoparticle material can be dissolved or dispersed in a solvent, then printed on a substrate, e.g. by spin coating, slit coating or doctor blading; vapour phase or evaporation techniques are far more expensive, requiring high temperatures and/or pressures. Secondly, nanoparticles are able to pack closely, facilitating their coalescence upon melting. Upon coalescence the particles can form large grains. Additionally, the melting point of nanoparticles is lower than that of the bulk material, allowing lower processing temperatures for device fabrication. Finally, nanoparticles can be synthesised in colloidal solutions. Colloidal nanoparticles may be capped with an organic ligand (capping agent); this assists in solubilising the particles, thus facilitating the processability of the material.
Nanoparticles can be synthesised from a top-down or a bottom-up approach. In a top-down approach, macroparticles are processed, e.g. using milling techniques, to form nanoparticles; the particles are typically insoluble, therefore difficult to process, and in the case of milling the size distribution may be large. Using a bottom-up approach, whereby nanoparticles are grown atom-by-atom, smaller particles with a homogeneous size distribution may be produced. Colloidal syntheses can be employed to grow nanoparticles in solution, which can be passivated with organic ligands to provide solubility, and thus solution processability.
The colloidal synthesis of CZTS nanoparticles has been described in the prior art. The colloidal synthesis of Cu2XSnY4 nanoparticles, where X is a d-block element and Y is a chalcogen, herein referred to as “CXTY”, is less well documented, however a number of examples exist.
Ou et al. described the synthesis of Cu2ZnSn(SxSe1-x)4 nanoparticles via the hot-injection of a solution of Cu, Sn and Zn stearate salts, dissolved in oleylamine, into a mixture of thiourea, oleylamine and octadecene, at 270° C. [K.-L. Ou et al., J. Mater. Chem., 2012, 22, 14667]
The hot-injection synthesis of Cu2ZnxSnySe4x+2y nanoparticles has been described by Shavel and co-workers. [A. Shavel et al., J. Am. Chem. Soc., 2010, 132, 4514] Trioctylphosphine selenide was injected into a solution of Cu, Zn and Sn salts dissolved in a mixture of hexadecylamine and octadecene, at 295° C.
The preparation of Cu2FeSnS4 nanoparticles has been described by Zhang et al. [X. Zhang et al., Chem. Commun., 2012, 48, 4656] A mixture of 1-dodecanethiol and t-dodecanethiol was injected into a solution of Cu, Fe and Sn salts in oleylamine at moderate temperature (150° C.). The solution was subsequently heated 210° C. to prepare wurtzite nanocrystals. To synthesise zinc blende nanocrystals, the solution was heated to 310° C., replacing oleylamine with oleic acid and octadecene.
The colloidal methods of making CXTY nanoparticle materials described in the prior art have one or more disadvantages including the use of hot-injection and/or high boiling capping agents (ligands).
Hot-injection techniques can be used to synthesis small nanoparticles with a uniform size distribution. The technique relies on the injection of small volumes of precursors into a large volume of solvent at elevated temperature. The high temperature causes breakdown of the precursors, initiating nucleation of the nanoparticles. However, the technique results in low reaction yields per volume of solvent, thus making the reactions difficult to scale to commercial volumes.
Other prior art techniques utilise high boiling ligands, such as oleylamine, hexadecylamine or oleic acid. Organic ligands assist in solubilising the nanoparticles to facilitate solution processability, yet they must be removed, e.g. by evaporation, prior to sintering, since residual carbon can be detrimental to the optoelectronic performance of the absorber layer. Thus it is favourable that the boiling temperature of any capping ligand(s) should be substantially lower than the sintering temperature of the CXTY film.
Thus, there is a need for a commercially scalable method to synthesise CXTY nanoparticle capped with a relatively low-boiling ligand that is suitable for low temperature optoelectronic device processing.